Astrocytes Protect Neurons against Methylmercury via ATP/P2Y1 Receptor-Mediated Pathways in Astrocytes

Methylmercury (MeHg) is a well known environmental pollutant that induces serious neuronal damage. Although MeHg readily crosses the blood-brain barrier, and should affect both neurons and glial cells, how it affects glia or neuron-to-glia interactions has received only limited attention. Here, we report that MeHg triggers ATP/P2Y1 receptor signals in astrocytes, thereby protecting neurons against MeHg via interleukin-6 (IL-6)-mediated pathways. MeHg increased several mRNAs in astrocytes, among which IL-6 was the highest. For this, ATP/P2Y1 receptor-mediated mechanisms were required because the IL-6 production was (i) inhibited by a P2Y1 receptor antagonist, MRS2179, (ii) abolished in astrocytes obtained from P2Y1 receptor-knockout mice, and (iii) mimicked by exogenously applied ATP. In addition, (iv) MeHg released ATP by exocytosis from astrocytes. As for the intracellular mechanisms responsible for IL-6 production, p38 MAP kinase was involved. MeHg-treated astrocyte-conditioned medium (ACM) showed neuro-protective effects against MeHg, which was blocked by anti-IL-6 antibody and was mimicked by the application of recombinant IL-6. As for the mechanism of neuro-protection by IL-6, an adenosine A1 receptor-mediated pathway in neurons seems to be involved. Taken together, when astrocytes sense MeHg, they release ATP that autostimulates P2Y1 receptors to upregulate IL-6, thereby leading to A1 receptor-mediated neuro-protection against MeHg.


Introduction
Methylmercury (MeHg), a well-known environmental pollutant, easily crosses the blood-brain barrier [1,2] inducing several types of serious neuronal damage and disorders [3,4,5,6]. Although most studies about MeHg-induced toxicity in the CNS have focused on its effects on neurons, MeHg, acting on a much higher number of glial cells, should affect their functions and viabilities. This is of great importance because it has become apparent that glial cells regulate a large variety of neuronal functions both in physiological and pathophysiological CNS [7]. However, the effects of MeHg on glial cells or neuron-to-glia interactions have received only limited attention.
Recently, it has become apparent that MeHg causes diverse responses in glial cells, i.e., it upregulates antioxidant genes [8,9], while it rather inhibits the uptake of cysteine, a critical precursor of glutathione synthesis, leading to a decrease in antioxidants [10]. As one of the mechanisms of MeHg-induced neuronal loss is oxidative stress [11,12,13,14], these glial responses by MeHg may greatly affect neuronal functions or viability. Inflammatory responses in glial cells are also involved in several types of neuronal damage. It has been reported that MeHg produces proinflammatory cytokines including interleukin-6 (IL-6) in glial cells [15,16,17]. In general, these cytokines facilitate inflammatory responses, leading to deterioration of the neuronal viability. However, we [18] and others [19] have already demonstrated that astrocytic IL-6 in response to various chemicals or insults protected neurons against oxidative neuronal death. However, the physiological or pathophysiological significance of the increased IL-6 in response to MeHg remains largely unknown, and even less is known about the mechanisms underlying MeHginduced IL-6 in astrocytes.
Here, we demonstrate that MeHg upregulates several genes in astrocytes, among which IL-6 is the highest. And, as mentioned above, astrocytes protect neurons against MeHg by IL-6-mediated mechanisms. We also demonstrate that, when astrocytes sense MeHg, they release ATP that autostimulates P2Y 1 receptors in astrocytes, thereby leading to IL-6 production via p38-mediated mechanisms. The released IL-6 appears to exhibit neuro-protection by upregulating adenosine A 1 receptors in neurons.

Chemicals and Antibodies
Reagents were obtained from the following sources. Adenosine 59-triphosphate (ATP), apyrase (grade III), bovine serum albumin Mice C57BL/6 mice (17-day-old fetal) were purchased from Japan SLC. P2Y 1 knock-out mice (C57BL/6 background) have been developed as previously reported [21]. The cortical astrocytes from these mice were prepared as it is for the rat cortical astrocytes.

WST-1 Assay
Neuronal viability and astrocytic viability were estimated by WST-1 assay using a cell counting kit (Dojindo, Kumamoto, Japan). After incubation with MeHg for 20 or 44 hr, 1/10 volume of WST-1 solution was added to the cell culture medium and incubated for an additional 4 hr. The absorbance of supernatants was measured with a microplate reader at 450 nm as the test wavelength and at 630 nm as the reference wavelength.

DNA Microarray Analysis
For this experiment, astrocytes were exposed to 10 mM of MeHg for 2 hr. Converting total RNA (100 ng) to the targets for Affymetrix GeneChip DNA microarray hybridization was done according to the manufacturer's instructions. The targets were hybridized onto a rat genome U34A GeneChip DNA microarray (Affymetrix, Santa Clara, CA) for 16-24 hr at 45uC. After hybridization, DNA microarrays were washed and stained on a Fluidics Station according to the protocol provided by Affymetrix. Afterward, the DNA microarrays were scanned, and then the images obtained were analyzed by GeneChip Operating System software (version 1.4; Affymetrix). The microarray data is available upon request.

Quantitative RT-PCR
Total RNA was isolated and purified from astrocytes and neurons using RNeasy (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Reverse transcription (RT)-PCR was performed using a one step primescriptH RT-PCR Kit (Takara Bio Inc., Shiga, Japan) according to the manufacturer's protocol. The reaction mix contained 40 ng of total RNA, 200 nM primers, 100 nM TaqMan probe, TAKARA EX TaqH HS and PrimeScript TM RT enzyme Mix. RT-PCR amplification and real-time detection were performed using an Applied Biosystems 7500 Real-Time PCR System (Applied Biosystems, CA, USA). The reverse transcription was performed at 42uC for 5 min followed by inactivation at 95uC for 10 s. The temperature profile consisted of 40 cycles of denaturation at 95uC for 5 s, and annealing/extension at 60uC for 34 s. The sequence of the primers and probe for rat IL-6 were as follows: the TaqMan probe, 59-CAGAATTGCCATTGCACAACTCTTTTCTCA-39; the forward primer, 59-CAGTGCATCATCGCTGTTCA-39; and the reverse primer, 59-CATATGTTCTCAGGGA-GATCTTGGA-39. The sequence of the primers and probe for rodent A 1 receptor were as follows: the TaqMan probe, 59-CGAGTCAAGATCCCTCTCCGGTACAAGA-39; the forward primer, 59-TCATCCTCACCCAGAGCTCC-39; and the reverse primer, 59-ATGGGTGTCAGGCCTACCAC-39. Primers and the Taquman probe for GAPDH were obtained from Rodent GAPDH Control Reagents (Applied Biosystems). Mouse IL-6 expression was estimated using the probe set (Mm0046190-m1) from Applied Biosystems (Foster City, CA).

Enzyme-linked Immunosorbent Assay of IL-6
The MeHg-induced IL-6 production from astrocytes was measured using a QuantikineH rat IL-6 immunoassay kit (R&D Systems, MN, USA). Astrocytes were incubated with MeHg (1 or 3 mM) in serum-free medium for 12 or 24 hr and the supernatants were collected. The assay was performed according to the manufacturer's instructions. All standards and samples were measured with a microplate reader at a wavelength of 450 nm.

Ca 2+ -imaging
Changes in intracellular Ca 2+ were measured by the fura 2 method with minor modifications [22]. In brief, the culture medium was replaced with balanced salt solution (BSS) of the following composition (in mM): NaCl 150, KCl 5.0, CaCl 2 1.8, MgCl 2 1.2, HEPES 25, and D-glucose 10 (pH 7.4). Cells were loaded with fura 2 by incubation with 10 mM fura 2-acetoxymethyl ester (fura 2-AM) at room temperature (RT) in BSS for 45 min. After loading, the samples were mounted on a microscope (ECLIPSE TE2000-U, Nikon,Tokyo, Japan) equipped with a 75-W xenon lamp and band-pass filters of 340 and 380 nm wavelengths for measurement of the Ca 2+ -dependent signals (F340 and F380 nm). Image data were recorded by a CCD camera (ORCA-ER, Hamamatsu Photonics, Shizuoka, Japan). For evaluation, we used the ratio of F340/F380.

Immunocytochemistry
Cells were fixed with 4% paraformaldehyde for 30 min at RT and they were incubated with the primary antibodies (anti-GFAP antibody at 1:2000; and anti-MAP2 antibody at 1:500) in a Can Get Signal A (TOYOBO, Osaka, Japan) for 24 hr at 4uC. Then, the cells were further incubated with Alexa 488-or Alexa 546conjugated second antibodies (1:2000) for 1 hr at RT. Fluorescent images were obtained by a laser scanning confocal microscope FV-1000 (Olympus, Tokyo, Japan).

Measurement of Extracellular ATP
The extracellular ATP concentration of the MeHg-treated astrocytes was determined with an ATP bioluminescence assay kit CLS II (Roche Applied Science, Mannheim, Germany). Astrocytes were incubated with MeHg (1 or 3 mM) in serum-free medium and the supernatants were collected and boiled at 95uC for 10 min. Equal volumes of luciferin/luciferase reagents and samples (100 ml each) were mixed a few times by gentle pipetting. All standards and samples were measured with a Lumat LB9501 tube luminometer (Berthold, Wildbad, Germany). The ATP concentrations were calculated from the intensities of a series of standard ATP.

Western Blotting
Cells were lysed and the lysates were electrophoresed with 10% SDS-PAGE gels and transferred to PVDF membranes. The membranes were blocked for 1 hr in Tris-buffered saline containing 0.1% Tween-20 and 5% BSA at RT and were incubated with primary antibodies (1:5000) over night at 4uC. Membranes were then incubated with horseradish peroxidase-conjugated 2nd antibodies (1:20000) for 1 hr at RT. Protein bands were visualized by rinsing the membrane with supersignal west pico chemiluminescence substrate (Thermo scientific, PA, USA). Images were obtained using LAS-4000 (Fujifilm, Tokyo, Japan).

Enzyme Histochemistry
Enzyme histochemistry for ecto-ATPases activity has been performed on the basis of a previous report [23]. Briefly, cells were fixed with 4% paraformaldehyde for 30 min at RT and preincubated for 30 min at RT with Tris-maleate-sucrose buffer (250 mM sucrose, 50 mM Tris-maleate, pH 7.4) containing 2 mM CaCl 2 . The enzyme reaction was performed in a reaction buffer (2 mM Pb(NO 3 ) 2 , 5 mM MnCl 2 , 2 mM CaCl 2 , 50 mM Tris-maleate (pH7.4), 250 mM sucrose, 3% dextran T250) with 1 mM of ATP as a substrate. After 1 hr reaction at RT, cells were washed with H 2 O and the ecto-ATPase activity was visualized by 0.5% (v/v) of (NH 4 ) 2 S.

Statistics
Data were expressed as means 6 SEM. Student's t-test was used for comparison of two groups. One way analysis of variance (ANOVA) followed by Tukey test was applied for multiple comparisons. The differences were considered to be significant when the P value was less than 5%.

MeHg Upregulates IL-6 Expression in Astrocytes
We first performed transcriptome analysis in cultured astrocytes stimulated with MeHg (10 mM) using DNA microarray (Table 1). MeHg changed the expressions of a large number of genes including those of cytokines and chaperones in astrocytes. Among them, interleukin-6 (IL-6) mRNA showed the most remarkable increase (638 fold), and we confirmed its upregulation using quantitative RT-PCR. The increase in IL-6 mRNA expression was concentration-dependent over a concentration range of from 0.1 to 3 mM with 2 hr-exposure (0.1 mM, 2.660.7; 1.0 mM, 6.160.7; 3.0 mM, 30.067.2 fold increase vs. control, n = 3) (Fig. 1A). The low concentration of MeHg (0.1 mM) never increased IL-6 mRNA expression at any exposure time tested

Discussion
MeHg easily passes the blood-brain barrier and causes serious damage in the CNS. Unlike neurons, its effect on glial cells has received only limited attention. In the present study, we demonstrated that, when exposed to MeHg, astrocytes exhibited neuro-protection against MeHg, in which ATP/P2Y 1 receptor-mediated signals and subsequent IL-6 production in astrocytes have a pivotal role. MeHg significantly increased extracellular ATP level of astrocytes. Although it has been described that MeHg induces astrocytic swelling [55,56], we did not observe significant morphological changes or injured structures in astrocytes when evaluated by immunocytochemical analysis using anti-GFAP ]i level in IL-6-treated (100 pg/ml, 24 hr) neurons was significantly lower than that in control neurons. This decrease was restored by DPCPX (1 mM). *P,0.05 vs. IL-6 alone. (ii) Representative traces of the glutamate-evoked increases in [Ca 2+ ]i in non-treated control (left), IL-6-treated (100 pg/ml, 24 hr) (right) and IL-6-treated neurons in the presence of 1 mM DPCPX. Glutamate (10 mM) was added to the neurons for 10 s. Bold line in each panel showed averaged changes in [Ca 2+ ]i in neurons, which was summarized in (iii). The glutamate-evoked increase in [Ca 2+ ]i in IL-6-treated neurons was significantly lower than that in control neurons, which was restored by DPCPX. **P,0.01 vs. Glu/IL-6. (D) A 1 receptor-mediated neuro-protection by IL-6. The protective effect of IL-6 (100 pg/ml) was suppressed by DPCPX (1 mM). **P,0.01 vs. MeHg/IL-6. (E) ATP-induced neuro-protection is mediated by A 1 receptor. Exogenously applied ATP (100 mM) restored the MeHg (1 mM, 48 hr)-reduced neuronal viability, and this effect was blocked by DPCPX (1 mM). *P,0.05, **P,0.01 vs. MeHg/ATP. (F) The MeHg-evoked increase in activity of ecto-ATPases in astrocytes. Activity of ecto-ATPases was analyzed by an enzyme histochemical assay. When stimulated with MeHg (3 mM), the activity (shown as brown signals) was increased, which peaked at around 6 to 12 hr. Scale bar, 200 mm. doi:10.1371/journal.pone.0057898.g006 Figure 7. A schematic diagram, illustrating mechanisms underlying astrocyte-mediated neuro-protection against MeHg. MeHg stimulates exocytosis of astrocytic ATP that functions as both (a) autocrine and (b) paracrine signals to reveal neuro-protection, i.e., (a) the released ATP as an autocrine signal, autostimulates P2Y 1 receptors to induce IL-6 that upregulates neuronal adenosine A 1 receptors, (b) the released ATP from astrocytes being degraded into adenosine, stimulates neuronal adenosine A 1 receptors and suppresses neuronal excitability as a paracrine signal, thereby leading to further inhibition of neuronal excitability. As for mechanisms for IL-6 synthesis and release, an increase in [Ca 2+ ]i in astrocytes mediated by P2Y 1 receptors, and subsequent p38 phosphorylation were involved (insert). doi:10.1371/journal.pone.0057898.g007 antibody. The increase in the extracellular ATP level did not seem to be due to leakage by cell damage because MeHg never decreased the astrocyte viability (Fig. 4). Although astrocytes release ATP by multiple mechanisms, exocytosis might be one of them because both BoNT and BAPTA-AM reduced the ATP release, while neither CBX nor Gd 3+ inhibited the release (Fig. 3). However, the finding that the inhibition of the Ca 2+ oscillation by BoNT was incomplete suggests that SNARE-independent pathways for ATP release might also be involved (Fig. 3F). With regard to ATP exocytosis, a recent report has demonstrated that lysosomes mediate at least part of the exocytotic ATP release in astrocytes [37]. In addition to lysosomes, vesicular nucleotide transporter (VNUT) has been reported to mediate exocytotic ATP release [57]. We must await further studies to clarify the involvement of lysosomes and/or VNUT-vesicles in the MeHgevoked ATP release by astrocytes or the initial target molecule(s) of MeHg. However, our present data clearly showed that, when astrocytes sense MeHg, they release ATP in part by exocytosis. This initial ATP release should be a key response because subsequent events, e.g., IL-6 production or neuro-protection by astrocytes, were dependent on ATP/P2Y 1 receptors.
In the transcriptome analysis, we found that MeHg upregulated several genes, among which IL-6 mRNA showed the most remarkable increase (table 1). The MeHg-evoked upregulation of IL-6 mRNA peaked at 2 hr and IL-6 protein production was observed at 12 or 24 hr ( Fig. 1A and 1B). Such time lag between IL-6 mRNA and protein expression can be also seen in other reports of various stimuli-induced IL-6 expression/production in astrocytes [58,59,60,61]. It would take longer time for the de novo synthesis or release of IL-6 after its mRNA upregulation in astrocytes, but we must await further investigation to clarify it since we did not check the IL-6 release at earlier (,12 hr) or later time points (.24 hr). The MeHg-evoked IL-6 production was dependent on the activation of P2Y 1 receptors in astrocytes, because the IL-6 production was (i) reduced by the antagonists to P2Y 1 receptors, (ii) was not observed in P2Y 1 R KO astrocytes, and (iii) was mimicked by exogenously applied ATP. As for the downstream signaling of P2Y 1 receptors, we focused on MAPKs. Of the three MAPK members (i.e. ERK1/2, JNK, and p38), only the p38 inhibitor SB203580 suppressed the MeHg-induced IL-6 mRNA upregulation (Fig. 2C). In addition, the phosphorylation of p38 by MeHg was blocked by the P2 receptor antagonist. ATP itself also evoked p38 phosphorylation (Fig. 2D). All these findings suggest that MeHg triggered the exocytosis of ATP, which in turn autostimulates P2Y 1 receptors, thereby leading to p38-mediated IL-6 production in astrocytes. Many reports have shown that p38 is required for the induction of IL-6 in astrocytes, together with a variety of stimuli including prostaglandin E 2 [62], co-stimulation with IL-6 and IL-17 [60], thromboxane A 2 [63], oncostatin M [64], and ICAM-1 ligation [65]. The p38 activation might be a common key pathway for IL-6 expression in astrocytes. The p38 inhibitor and P2 receptor antagonists (i.e. suramin and MRS2179) exhibited lesser extent of inhibitory effects than those by P2Y 1 R KO astrocytes. This discrepancy might be due to their lower concentrations because previous studies have shown that 10 mM of SB203580 does not show complete blockade for stimuli-induced p38 phosphorylation in astrocytes [66,67,68,69,70]. Similarly, we previously showed that neither suramin (100 mM) nor MRS2179 (1 mM) completely suppressed the ATP (100 mM)-evoked Ca 2+ transient (i.e. about 70% suppression) [26].
Since IL-6 is a cytokine with major regulating effects on the inflammatory response, in general an elevation in the proinflammatory cytokine IL-6 is considered to have a damaging effect on neurons. We and others, however, reported that IL-6 could protect neurons against a variety of damages including trauma, ischemia, excitotoxicity and oxidative stress [18,19,46,47,48,49,50,51,52]. In the present study, astrocytes showed neuro-protection against MeHg via IL-6-mediated mechanisms because ACM-induced neuro-protection was IL-6 dependent (Fig. 5) and exogenously applied recombinant IL-6 protein mimicked the neuro-protection (Fig. 4). One possible mechanism of the IL-6-mediated neuro-protection is that IL-6 stimulates the induction of neuro-protective molecules. We showed that the neuro-protection by IL-6 disappeared in the presence of a protein synthesis inhibitor CHX (Fig. 6A), suggesting that de novo synthesis of certain neuro-protective molecules appears to be required. Recent reports by Biber et al. demonstrated that IL-6 inhibits the glutamate-induced excitototoxicity of cortical neurons requires adenosine A 1 receptor functions in neurons [53,54]. The IL-6 increased both mRNA and proteins of adenosine A 1 receptors in the neurons, and the protection by IL-6 disappeared in the presence of CHX [54]. In the present study, we also found that IL-6 increases A 1 receptor mRNA expression in cortical neurons (Fig. 6B). IL-6 increased not only mRNA but also A 1 receptor-mediated tonic inhibition on an excitatory neurotransmitter (Fig. 6C). Supporting these results, the IL-6-induced neuro-protection against MeHg was suppressed by the A 1 receptor antagonist DPCPX (Fig. 6D). All these findings may support the idea that one of the neuro-protective molecules induced by ACM or IL-6 would be adenosine A 1 receptors.
However, anti-IL-6 antibody could not abolish the effect of ACM, indicating the involvement of IL-6-independent neuroprotective mechanisms. We considered that the astrocyte-derived ATP itself might function as another neuro-protective molecule because, without ACM, the exogenously applied ATP alone showed neuro-protection (Fig. 6E). Interestingly, this protection by ATP was also inhibited by the antagonist to adenosine A 1 receptor DPCPX (Fig. 6E). Astrocytic ATP either as ATP [20] or metabolized into adenosine by ecto-nucleotidases [71,72], inhibits excess excitatory synaptic transmission, leading to inhibition of excitatory neuronal death. Our time-lapse analysis of extracellular ATP level and enzyme histochemistry have shown that MeHg gradually increased ATP release from astrocytes followed by an increase in ATPase activity ( Fig. 3B and 6F). Under this condition, extracellular adenosine would increase and activate neuronal A 1 receptor.
Since it takes 24 hr for IL-6 to protect neurons against MeHg (Fig. 4), the delayed production of IL-6 ( Fig. 1) in astrocytes might be problematic to the neuro-protection, because in general both neurons and astrocytes would be simultaneously exposed to MeHg in situ. However, astrocytes could release ATP in response to MeHg as early as 15 min after MeHg (Fig. 3A), and the released ATP or its metabolite adenosine directly protected neurons against MeHg (Fig. 6), suggesting that astrocytes should show the IL-6independent neuro-protection even in the early stage. Thus, the astrocyte-mediated neuro-protection shown in the present study could work in situ. Furthermore, we previously showed that activation of P2Y 1 receptors in astrocytes increased tolerance against oxidative stress by the upregulation of various oxidoreductase genes [26,73]. Therefore, the astrocytic ATP release and activation of P2Y 1 receptors appear to be key events that trigger multiple neuro-protective responses.
Taken together, as summarized in Figure 7, when astrocytes are exposed to MeHg, they exocytose ATP and show a neuroprotective phenotype. The released ATP functions as an autocrine to stimulate P2Y 1 receptors, thereby leading to the protection of neurons against MeHg via IL-6-mediated pathways. The IL-6 increases neuronal A 1 receptor expression and function. The released ATP, being metabolized into adenosine, may also function as a paracrine to exert neuro-protection via suppressing excitatory neurotransmission. Figure S1 Differences in Ca 2+ responses to 2MeSADP and UTP in WT and P2Y 1 R KO mice. (A) Typical Ca 2+ responses to the P2Y 1 R agonist 2methyl-thio-ADP (2MeSADP) (1 mM) and the P2Y 2/4 receptor agonist UTP (100 mM) in control astrocytes obtained from WT mice (upper traces) and those from P2Y 1 R KO mice (lower traces). Although UTP evoked [Ca 2+ ]i elevations in both WT and P2Y 1 R KO astrocytes, 2MeSADP failed to produce the [Ca 2+ ]i increse in P2Y 1 R KO astrocytes, which was summarized in B. (TIF)